[rohc] RE: Reveiw - draft-ietf-rohc-rfc3095bis-framework-02.txt

Carl,

Thanks for reviewing the document, please see responses inline.

BR
/L-E

----Original Message----
From: Carl Knutsson [mailto:carl.knutsson@effnet.com]
Sent: den 1 november 2006 14:08
To: rohc@ietf.org
Cc: Lars-Erik Jonsson (LU/EAB)
Subject: Reveiw - draft-ietf-rohc-rfc3095bis-framework-02.txt

> Hi L-E, all,
> 
> I been asked to review the framework draft.
> 
> My comments:
> 
> 1. Introduction
> 
> "To improve and simplify the specification, the framework and
> the profiles parts of RFC 3095 have been split into separate
> documents."
> 
> The sentence above could be rephrased without RFC 3095.
> 
> "To improve and simplify the specification, the framework and
> the profiles parts have been split into separate documents."
> 
> We are splitting the new specification into two documents.
> This is a new specification. The sentence above gives the impression
> that we have only split the rfc 3095 into two new documents.

Agree!

> 2.2. ROHC Terminology
> 
> Some of the definitions in the Terminology section are not
> used very frequently.
> 
> Context Repair mechanism - only used only in history part. "Repair
>                            mechanism" is used. Maybe it should be
>                            added in terminology.

I think Context Repair Mechanism is the correct term to define, although
the text sometimes use a shorter form. In section 6, I suggest we add
"context" to the last sentence to make it "The profile defines context
repair mechanisms, including feedback logic if used."

> CRC-8 validation - not used in document. (Could be used in
> section 5.4.3)
> 
> CRC verification - not used in document.

These are conceptual ROHC terms, which may be used by profile
specifications. Maybe we can use these in 5.2.3 and 5.3, but not
necessarily.

> Error propagation - not used in this document.

True, but it is a conceptual term (used before), and it may well be
referred to by profiles.

> Packet Flow - Not used in this document. "flow of packet" and "flow"
> is used. Include "flow" in the terminology?

I think Packet Flow is the correct term to define, although the text
sometimes use a shorter form. We should make sure the most important
occurencies are changed to say "packet flow".

> Header Compression Profile - not used in document. "ROHC profile" and
> "profile" is frequently used in the draft.

I think the most accurate term for the terminology section would be
"ROHC Profile", I'll change this.

> 5.1.1. Contexts and Context Identifiers.
> 
> "..or when the CID gets associated from the reception of an
> IR (this does not apply to the IR-DYN) with a different profile
> than the profile in the context."
> 
> IR and IR-DYN mentioned the first time (except for Abbreviation and
> Terminology section). I suggest adding a reference to section 5.2.2.

I looked at this and it easily gets messy. IR and IR-DYN are defined
in the acronyms section, that should be enough in my opinion.

> 5.2.2. Initialisation and Refresh (IR) Packet Types
> 
> last sentence last paragraph...
> 
> ".., or unless the profile in the Profile field specifies otherwise"
> 
> could be rephrase to
> 
> ".., or unless the profile indicated in the Profile field specifies
> otherwise"

Agree!

/L-E

_______________________________________________
Rohc mailing list
Rohc@ietf.org
https://www1.ietf.org/mailman/listinfo/rohc